AITRICS, a Korean medical AI company, said Tuesday that its research on deep learning models for predicting the risk of acute kidney injury (AKI) has been published in npj Digital Medicine. The study developed and evaluated multiple AI models designed to predict the onset of AKI and compared their performance using an online simulation framework that generated predictions every 12 hours to mimic continuous clinical monitoring. The AITRICS team developed three deep learning models and two conventional machine learning models using data from 157,323 hospitalized patients across three hospitals in Korea and abroad. The results showed that the deep learning models outperformed the conventional machine learning models in predictive performance. However, the model that achieved the highest accuracy at a single time point did not necessarily deliver the best performance under real clinical conditions. The study found that evaluating models based on both the stability of predictive performance over time and alert burden enabled researchers to identify models better suited for clinical implementation. The researchers said that incorporating assessments of temporal model stability and alert burden provides a more practical framework for selecting AI models suitable for routine clinical use.